wined3d: declaration/FVF conversion test

[Ivan Gyurdiev]

Stefan Dösinger wrote:
>> and those currently fail, because that case isn't handled at all in the
>> FVF loading code. It thinks this is a multi-stream case, when in fact
>> only the 0th stream is used. Jason Green has a demo -
>> dx9_hdr_texture_loader, which is broken exactly by this.
>>     
> How about getting rid about the fvf in the device entirely and converting the 
> fvf to a vertex decl in SetFVF?
That's what I was suggesting in the next sentence of that email :)
>  This would make the drawprim code 
> simpler(only vertex declaration to consider) and avoid inconsistency. In 
> GetFVF we convert the declaration back to a FVF(and propably cache the 
> result, but don't use that for rendering).
>   
Well, caching is already accomplished by storing the FVF [ so no need 
for backwards conversion]. Furthermore, the backwards conversion isn't 
very clear at the moment - the tests show that almost all FVF fields are 
0'ed when a declaration is implicitly converted to FVF by windows. I 
have no idea why that is yet...